Typically, an automatic (non-human) speech recognition/transcription (AST) converts speech to text via a single speech recognition model. Typically, such single speech recognition model only transcribes a single language using a single acoustic model and is based on trade-offs in accuracy between general and specific language patterns.